Compaction-induced stress: Numerical

Modeling

S.H. MIRMORADI1 and M. EHRLICH

Federal University of Rio de Janeiro, UFRJ, RJ 21945-970, Brazil

Abstract. Two numerical procedures for modeling of compaction-induced stress

used in the literature were described, and analyses using those procedures were
performed The value of the maximum tension in the reinforcements determined
using both numerical procedures were compared with those measured from a
wrapped-faced full-scale reinforced soil wall performed at the Geotechnical
Laboratory of COPPE/UFRJ. Results show that simulation of the compaction-
induced stress by applying an equal distribution load at the top and bottom of each
soil layer was able to properly represent the measured values of the summation of
the maximum tension in the reinforcement. It was verified that modeling of
compaction using a uniform vertical stress applied only to the top of each backfill
soil layer over estimated the measured values, and this discrepancy increases with
depth. Furthermore, parametric studies show that the tension in the reinforcement
is a function of the magnitude of the compaction-induced stress. Nevertheless, the
induced stress due to compaction may not play an effective role in the magnitude
of the horizontal toe load. Note that in the analyses, the backfill soil parameters
were kept constant. However under actual field conditions, variations in these
parameters may influence results.

Keywords. Geosynthetic; GRS walls; Compaction-induced stress; Numerical

modeling; AASHTO simplified method; Ehrlich and Mitchell (1994) method.

1. Introduction

The importance of backfill soil compaction on the behavior of geosynthetic reinforced

soil (GRS) walls has been demonstrated by a number of laboratory and case studies
(Ehrlich and Mitchell 1994; Bathurst et al. 2009; Ehrlich et al. 2012). In recent decades,
several numerical analyses using the finite element method (FEM) or finite difference
method (FDM) codes have been undertaken to consider the different geometries and
parameters in GRS walls. However, the effect of backfill soil compaction has rarely
been considered.
Hatami and Bathurst (2005) and Guler et al. (2007) numerically considered the
effect of the induced stress due to backfill compaction using FDM (finite difference-
based fast Lagrangian analysis of continua program; Itasca Consulting Group, 2001)
and FEM (PLAXIS). In both studies, the compaction-induced stresses were modeled
using a uniform vertical stress applied only to the top of each backfill soil layer as the
wall was modeled from the bottom up. Ehrlich and Mirmoradi (2013), Mirmoradi and
Ehrlich (2014a,b) and Ricco et al. (2014) and Mirmoradi and Ehrlich (2015) simulated

1
Corresponding Author.
the compaction-induced stress by applying an equal distribution load at the top and
bottom of each soil layer. The difference between the two approaches on representation
of the actual field behavior are presented and discussed.

2. Modeling of Compaction-induced stress

Figure 1 shows a schematic view of the vertical stress increase during a roller operation
in backfill. The vertical stress at the top of each layer during the compaction roller
operation may be represented by a strip load and an elastic solution can be used to
represent its evolution with depth. As shown in figure 1 for each soil layer, the
maximum stress increase during the roller operation occurs at soil-roller contact and
decreases with depth. Ehrlich and Mitchell (1994) stated that “In multilayer
construction, the compacted layers are relatively thin, typically 0.15–0.3 m thick, and
all points in each soil layer may be assumed to have been driven to the same maximum
soil stress state during compaction”. Therefore, it may be assumed that all points in the
soil layers are driven to the same vertical induced stress, σ'zc,i, due to compaction.

Figure 1. Vertical stress increase during a roller operation in the backfill soil (strip load - Boussinesq elastic
solution).

Figure 2 shows two different approaches for the simulation of the induced stress
due to compaction. Figures 2a and 2b show a schematic view of the numerical
modeling of compaction-induced stress using a distributed load, qc, at the top of each
soil layer (hereafter referred to as procedure type 1) and distribution loads, qc, at the top
and bottom of each soil layer (hereafter referred to as procedure type 2), respectively.
Stage construction is used in both procedures and compaction modeling is represented
by only one cycle of loading and unloading for each soil layer. In Figure 2, four steps
for backfill soil construction in a specific soil layer "n" were considered: (I) soil layer
placement, (II) compaction equipment operation, (III) end of compaction, and (IV) next
soil layer placement (layer "n+1"). Figure 2a, step (II) shows that when procedure type
1 is used for numerical modeling of the induced stresses due to compaction in soil layer
“n”, it leads to the constant increase of the vertical stress due to compaction, qc, in all
the layers placed bellow. The dashed line in this figure shows the expected vertical
stress increase during the roller operation for soil layer “n” based on the strip load
elastic solution, where its maximum value takes place at soil-roller contact and
decreases significantly with depth. This figure clearly shows that using the distribution
load only at the top of each soil layer for modeling of compaction may not correspond
to the actual field condition represented by the elastic solution.
 Figure 2b shows a schematic view of the procedure type 2 as suggested by Ehrlich
and Mirmoradi (2013), Mirmoradi and Ehrlich (2014a,b) and Mirmoradi and Ehrlich
(2015) for the numerical simulation of the induced stress due to compaction. Figure 2b,
step (II) shows that when procedure type 2 is used for the soil layer “n”, all points in
this soil layer would be driven to the same vertical stress increase. In addition, for the
soil layers placed under this layer, only geostatic stresses occur. A comparison between
the curves related to the compaction modeling using procedure type 2 and the dashed
line represented by the elastic solution indicates that this procedure may be more
representative of the actual induced vertical stress during the roller operation.

(a)

(b)
Figure 2. Modeling of the vertical stress load-unload cycles verified during the compaction of the backfill
soil layer "n": a: procedure type 1, b: procedure type 2.

3. Numerical Modeling

The numerical modeling was carried out using the two-dimensional finite-element
program PLAXIS (Brinkgreve and Vermeer, 2002). Full-scale reinforced soil wall
modeling performed at the Geotechnical Laboratory of COPPE/UFRJ was used for the
validation of the numerical analyses. Details of the physical model and validation of
the numerical modeling can be found in Ehrlich and Mirmoradi (2013). Table 1 lists
the input parameters used in this study.
In order to verify the numerical modeling of compaction induced stress in GRS
walls using the procedures described above, the measured values of the summation of
the maximum tension mobilized in the reinforcement, ∑Tmax, was compared with those
determined by PLAXIS for compaction procedures type 1 and 2, the AASHTO
simplified method and the values predicted by the Ehrlich and Mitchell (1994) methods
in Figure 3. The vertical dotted line in this figure represents the compaction influence
depth, Zc defined as follows (Ehrlich and Mitchell 1994):
  zc ,i
Zc  (1)


where σ'zc,i and γ are the vertical stresses induced during compaction and the soil
unit weight, respectively. The equivalent depth of the soil layer (Zeq) is defined by:

q
Z eq  Z  (2)


where Z and q are the real depth of a specific layer and the surcharge load value of
the physical model, respectively.

Table 1. Input parameters for numerical analysis.

Property Value
Soil
Model HS
Peak plane-strain friction angle, φ (°) 50
Cohesion, c (kPa) 1
Dilation angle, Ψ (°) 0
Unit weight, γ (kN/m3) 21
E50ref (kPa) 42500
Eoedref (kPa) 31800
Eurref (kPa) 127500
Stress dependence exponent, m 0.5
Failure ratio, Rf 0.7
Poisson’s ratio, υ 0.25
Reinforcement
Elastic axial stiffness (kN/m) 600
Wrapped face
Elastic axial stiffness (kN/m) 60
Elastic bending stiffness (kNm2/m) 1

As shown in Figure 3, the values measured from the physical model were properly
captured by the Ehrlich and Mitchell 1994 method, and the numerical modeling using
compaction procedure type 2. However, regardless of the value of Zeq, the curve
corresponding to the numerical simulation using compaction procedure type 1
overestimates the values of ∑Tmax and this discrepancy increases with equivalent depth.
Comparison of the results corresponding to the conditions with and without
induced stresses due to compaction illustrates that for compacted backfill soil walls,
when Zeq<Zc, the values of ∑Tmax are greater than the values obtained for the no-
compaction conditions. However, for Zeq>Zc, the compaction-induced stress was
overcome by the geostatic stress and the values determined are the same irrespective of
whether or not the induced stress due to the backfill soil compaction is considered in
the analysis.
In Figure 3, the results related to the condition without compaction are also shown.
These curves were obtained with the Ehrlich and Mitchell (1994) method, the
AASHTO simplified method, and by numerical modeling with PLAXIS. The results
show similar values for the no-compaction condition. However, the AASHTO
simplified method leads to lower ∑Tmax values due to the assumption that the value of
the lateral earth pressure coefficient is taken to be equal to Ka.

32
Measured Zc
AASHTO
E&M, NC
E&M, 63 kPa
24
Summation of Tmax, kN/m

PLAXIS, NC
PLAXIS, Type 2
PLAXIS, Type 1

16

0
0 1 2 3 4 5 6
Equivalent depth, m
Figure 3. Comparison of measured and determined summations of the maximum tension along the
reinforcement layers.

4. Parametric Study

Parametric studies were carried out with different combinations of wall height, backfill
soil compaction efforts and facing stiffness. Three different wall heights, H, were
considered: 4 m, 8 m, and 16 m. The length and the vertical spacing of reinforcements
were 0.7H and 0.4 m, respectively. The block face was vertical. Figure 4 shows the
geometry of the numerical model for a wall with a height of 4 m. The facing stiffness
was evaluated by considering a block facing with a different stiffness modulus. The
values of facing stiffness were calculated using the different magnitudes of the stiffness
modulus multiplied by the area moment of inertia of the solid block. Four different
compaction effort factors were considered (0.0, 0.5, 1.0, and 2.0). The compaction
effort factor, CF, is defined as follows:

 'zc ,i
CF  (3)
H

where σ , is the vertical stress induced during backfill compaction, γ is the soil
unit weight, and H is the wall height. Table 2 shows the input parameters used for
modular block facing and the interfaces.
 0.4m

Blcok
0.4m
H=4m
Backfill

Reinforcement

2.8 m
B=6 m

Figure 4. Typical numerical model.

Figures 5 and 6 show the normalized summation of the maximum tension in the
reinforcements, and the normalized horizontal toe load, versus the normalized facing
stiffness. The results are related to three different wall heights (4 m, 8 m, and 16 m),
and four different values of CF (CF= 0.0, 0.5, 1.0, and 2.0). In these figures, the
symbols represent the different values of CF and the wall heights. Figures 5 and 6 show
that an increase in the facing stiffness results in a reduction of the maximum tension in
the reinforcements and a greater horizontal toe load. This is in agreement with the
results presented by Tatsuoka (1993), Bathurst et al. (2006), and Ehrlich and
Mirmoradi (2013).

Table 2. Modular block and interfaces properties.

Property Value
Modular block properties

Model Linear elastic

Size, (m×m) 0.4×0.2 (long×high)
Unit weight, γ, (kN/m3) 22
Stiffness modulus, (kPa) 1 × 106, 5 × 106, 1 × 107, 5 × 107
Poisson’s ratio, υ 0.15

Block-block interface
Friction angle, (0) 57
Cohesion, (kPa) 46

Soil-block interface

Friction angle, (0) 44

Cohesion, (kPa) 1
Dilation angle, Ψ, (o) 11

Furthermore, these figures indicate that a decrease of wall height H leads to a

lower value of the maximum tension in the reinforcements due to an increase in the
normalized horizontal toe load. This means that the wall height should be taken into
consideration for representing the facing stiffness. Note that the parameter used to
represent facing stiffness is simplistic. Besides the wall height, reinforcement spacing
and the shear stiffness of the interface between blocks may have a significant effect on
results. Nevertheless, comparison of the results is valid because these factors were
assumed to be the same in all of the analyses. For a larger compaction effort factor CF,
the maximum tension in the reinforcement was larger. However, the normalized
horizontal toe loads showed no significant variation with an increase in the compaction
effort factor, CF. Note that the objective of those analyses was to verify fundamental
aspects of the effect of σ , on the behavior of reinforced soil walls. In the analyses, the
backfill soil parameters (γ, φ, E , etc) were the same regardless of the value of σ , .
However, backfill soil parameters are dependent on compaction, and under actual field
conditions, variations in these parameters may influence results.

0.06
0.5

1.0
0.05
CF=0.0 2.0

0.04

∑T max
0.03
γH
2

0.02     CF   0.0     0.5      1.0      2.0 
 H  
0.01  4  m 
 8  m
16 m 
0.00
1E‐4 1E‐3 0.01 0.1 1 10 100
EI
γH
5

Figure 5. The normalized summation of the maximum tension in the reinforcements versus the normalized
facing stiffness for different values of the wall height and compaction effort factor, CF.

0.10

0.08

0.06
F
γ H2 0.04
    CF   0.0      0.5      1.0      2.0 
 H  
 4  m 
0.02
 8  m
16 m 

0.00
1E‐4 1E‐3 0.01 0.1 1 10 100
EI
γH
5

Figure 6. The normalized horizontal toe load versus the normalized facing stiffness for different values of
the wall height and compaction effort factor, CF.
5. Conclusions

Two numerical procedures for modeling of compaction were described, and analyses
using those procedures were performed to evaluate the effect of the induced-stresses
due to compaction on the behavior of GRS walls. Results show that modeling of
compaction induced stress using two distributed loads at the top and bottom of each
soil layer (type 2) were able to properly represent the measured values and simulate the
effect of compaction observed in the physical model studies. On the other hand
modeling of compaction using a distribution load only at the top of each soil layer (type
1) overestimated the measurements, and the discrepancy increased with depth.
Furthermore, parametric study using the compaction procedure type 2 shows that
the tension in the reinforcement is a function of the magnitude of the compaction-
induced stress. Nevertheless, the induced stress due to compaction may not play an
effective role in the magnitude of the horizontal toe load. Note that in the analyses, the
backfill soil parameters were kept constant. However under actual field conditions,
variations in these parameters may influence results.

References

[1] AASHTO, AASHTO LRFD Bridge Design Specifications, American Association of State Highway and
Transportation Officials, Sixth Edition, Washington, D.C., USA, 2012.
[2] Bathurst, R.J., Nernheim, A., Walters, D.L., Allen, T.M., Burgess, P., Saunders, D.D., Influence of
reinforcement stiffness and compaction on the performance of four geosynthetic reinforced soil walls.
Geosynth. Int. 16 (1), (2009), 43–59.
[3] Bathurst, R.J., Vlachopoulos, N., Walters, D.L., Burgess, P.G., Allen, T.M.,The influence of facing
stiffness on the performance of two geosynthetic reinforced soil retaining walls. Can. Geotech. J. 43,
(2006), 1225-1237.
[4] Brinkgreve, RBJ., Vermeer, PA., PLAXIS: finite element code for soil and rock analyses. Version 8.
Balkema, 2002.
[5] Ehrlich, M., Mirmoradi, S.H., Saramago, R.P., Evaluation of the effect of compaction on the behavior of
geosynthetic-reinforced soil walls. Geotextiles and Geomembranes 34, (2012), 108-115.
[6] Ehrlich, M., Mirmoradi, S.H., Evaluation of the Effects of Facing Stiffness and Toe Resistance on the
Behavior of GRS Walls. Geotextiles and Geomembranes 40, (2013), 28-36.
[7] Ehrlich, M., Mitchell, J. K., Working stress design method for reinforced soil walls. Journal of Geot.
Eng., ASCE, 120(4), (1994), 625–645.
[8] Guler, E., Hamderi, M.,Demirkan, M. M., Numerical analysis of reinforced soil-retaining wall structures
with cohesive and granular backfills. Geosynthetics International 14 (6), (2007), 330-345.
[9] Hatami, K., Bathurst, R. J., Development and verification of a numerical model for the analysis of
geosynthetic reinforced-soil segmental walls. Can. Geotech. J., 42 (4), (2005), 1066–1085.
[10] Mirmoradi, S.H., Ehrlich, M., Modeling of the Compaction-induced Stresses in Numerical Analyses of
GRS Walls. International Journal of Computational Methods (IJCM), Special Issue on "Computational
Geomechanics", 11(2), (2014a), 14 p.
[11] Mirmoradi, S.H., Ehrlich, M., Numerical Evaluation of the Behavior of GRS Walls with Segmental
Block Facing under Working Stress Conditions. ASCE J. Geotech. Geoenviron. Eng., (2014b), (DOI:
10.1061/(ASCE)GT.1943-5606.0001235).
[12] Mirmoradi, S.H., Ehrlich, M., Modeling of the compaction-induced stress on reinforced soil walls.
Geotextiles and Geomembranes 43 (1), (2015), 82-88.
[13] Riccio, M., Ehrlich, M., & Dias, D.,Field Monitoring and Analyses of the Response of a Block-faced
Geogrid Wall using Fine-grained Tropical Soils. Geotextiles and Geomembranes, 42(2), (2014), 127-
138.
[14] Tatsuoka, F., Roles of facing rigidity in soil reinforcing. Keynote Lecture. Proc. Earth Reinforcement
Practice, Fukuoka, Japan, 2, (1993), 831-870.